Skip to content

Free tool

LLM cost calculator

Enter your monthly request volume and average token lengths to see the real cost difference between models — cache hit rate included.

Your usage profile

200,000
4,000
600
%60

Share of repeated system prompt and context served from cache.

Prices are USD per million tokens, taken from each provider's own pricing page. Indicative only — not a quote. Per-provider cache discounts are included; cache-write cost and promotional or context-tier differences are not.
Price data: 2026-08-18

Estimated monthly cost

  • Mistral Small 4cheapest

    Mistral · $0.15/$0.6 · $0.015

    $127

    Monthly · $0.0006 per request

  • GPT-5.6 Luna

    OpenAI · $0.2/$1.2 · $0.02

    $218

    Monthly · $0.0011 per request · 1.7× vs. cheapest

  • DeepSeek V4 Flash

    DeepSeek · $0.44/$1.32 · $0.014

    $306

    Monthly · $0.0015 per request · 2.4× vs. cheapest

    Peak-hour rate; a cache hit takes input cost close to zero.

  • Mistral Large 3

    Mistral · $0.5/$1.5 · $0.05

    $364

    Monthly · $0.0018 per request · 2.9× vs. cheapest

  • Gemini 3.5 Flash-Lite

    Google · $0.3/$2.5 · $0.03

    $410

    Monthly · $0.0021 per request · 3.2× vs. cheapest

  • Gemini 3.7 Flash

    Google · $0.75/$3.75 · $0.075

    $726

    Monthly · $0.0036 per request · 5.7× vs. cheapest

    Introductory pricing through 31 Dec 2026; it doubles afterwards.

  • Grok 4.3

    xAI · $1.25/$2.5 · $0.2

    $796

    Monthly · $0.004 per request · 6.3× vs. cheapest

  • GPT-5.4 mini

    OpenAI · $0.75/$4.5 · $0.075

    $816

    Monthly · $0.0041 per request · 6.4× vs. cheapest

  • DeepSeek V4 Pro

    DeepSeek · $1.32/$3.96 · $0.044

    $919

    Monthly · $0.0046 per request · 7.2× vs. cheapest

    Peak-hour rate; off-peak hours are charged at half.

  • Claude Haiku 4.5

    Anthropic · $1/$5 · $0.1

    $968

    Monthly · $0.0048 per request · 7.6× vs. cheapest

  • Mistral Medium 3.5

    Mistral · $1.5/$7.5 · $0.15

    $1,452

    Monthly · $0.0073 per request · 11.4× vs. cheapest

  • Grok 4.6

    xAI · $2/$6 · $0.5

    $1,600

    Monthly · $0.008 per request · 12.6× vs. cheapest

    Pricing doubles above a 200k-token prompt.

  • Gemini 3.5 Flash

    Google · $1.5/$9 · $0.15

    $1,632

    Monthly · $0.0082 per request · 12.8× vs. cheapest

  • GPT-5.6 Terra

    OpenAI · $2/$12 · $0.2

    $2,176

    Monthly · $0.0109 per request · 17.1× vs. cheapest

  • Gemini 3.1 Pro

    Google · $2/$12 · $0.2

    $2,176

    Monthly · $0.0109 per request · 17.1× vs. cheapest

    Input and output rates rise above a 200k-token prompt.

  • Claude Sonnet 5

    Anthropic · $3/$15 · $0.3

    $2,904

    Monthly · $0.0145 per request · 22.8× vs. cheapest

    The best speed/intelligence balance; near-Opus on coding and agentic work.

  • Claude Opus 5

    Anthropic · $5/$25 · $0.5

    $4,840

    Monthly · $0.0242 per request · 38.1× vs. cheapest

    The default for complex agentic coding and enterprise work.

  • GPT-5.6 Sol

    OpenAI · $5/$30 · $0.5

    $5,440

    Monthly · $0.0272 per request · 42.8× vs. cheapest

  • Claude Fable 5

    Anthropic · $10/$50 · $1

    $9,680

    Monthly · $0.0484 per request · 76.1× vs. cheapest

    For the hardest reasoning and long-horizon agentic work.

  • GPT-5.5 Pro

    OpenAI · $30/$180 · no cache discount

    $45,600

    Monthly · $0.228 per request · 358.5× vs. cheapest

    No cache discount — cost scales fast on long-context workloads.

Let's bring the cost down with architecture

Model choice is only part of the cost. Cache strategy, context length and routing architecture usually make a bigger difference.